Posts by Eric B

1) Message boards : Cafe SETI : Question on non-seti project (Message 2075958)
Posted 16 May 2021 by Profile Eric B
Post:
When seti quit, I switched over to milkyway at home and crunched over there since. About two or three weeks or so ago they vanished, no home page, no message boards, no work, and no way to report work. Its like they just dropped off the edge of the earth.
Does anyone know what happened there?
2) Message boards : Number crunching : openCL issue (Message 2053010)
Posted 1 Jul 2020 by Profile Eric B
Post:
I have two GPU's (NVIDIA RTX 2060 and a NVIDIA GTX 1660 ti under openSuse Tumbleweed Linux ) that i used to run seti and it worked very well, now i am running Milkyway at home and that will still use the GPU's
I had been trying to run 1 monitor per gpu but could never gt it working, so I reconnected my monitors to a single GPU and got both my screens working, But there is another issue.
On boot ,as a regular user, clinfo shows no platforms, but if i then run clinfo as root it shows both GPU's ok, and after that if I again run clinfo as a regular user it shows shows both GPU's.
This is repeatable and happens at every boot.
So, running clinfo as root changes something, i dont know what tho.
If i dont follow the above procedure boinc doesnt see either GPU.
If anyone can explain this I'd be deeply appreciative.
TIA
3) Message boards : News : SETI@home hibernation (Message 2042193)
Posted 1 Apr 2020 by Profile Eric B
Post:
Well, over the years I managed to accumulate 167,563,032 credits, I sure wish we had found something.
4) Message boards : Number crunching : The Server Issues / Outages Thread - Panic Mode On! (118) (Message 2033304)
Posted 21 Feb 2020 by Profile Eric B
Post:
I'm afraid this is going to recur over and over until they reach out to someone (google for example) who deals with very large and very busy databases, for help. No amount of shutting down and letting things catch up is going to solve this problem. As far as I know they haven't even been able to root cause the issue.
5) Message boards : Number crunching : Need Help: trying to get 2 nvida cards runnng under linux (Message 2032191)
Posted 13 Feb 2020 by Profile Eric B
Post:
I found it, something disabled my affiliation with 'video' users, once i put that back everything worked
My boinc install is a bit different. I dont use the boinc user/group because i like to run boinc-client as
myself in $HOME/boinc and in the foreground so i can see whats going on there. Another reason is the
default suse install never worked properly and no matter what i tried it always tried to start boinc in the
wrong place even with the -d or --dir switch which it mostly ignored. There were other reasons as well, but
now that behind me and i have been running boinc-client i the foreground and in $HOME/boinc on 4
machines without issue until today when i added the 2060. 7 days ago there was a broken update from
suse that may have been the root cause of the group change - whatever, its fixed now.
 Starting BOINC client version 7.16.3 for x86_64-suse-linux-gnu
 log flags: file_xfer, sched_ops, task
 Libraries: libcurl/7.68.0 OpenSSL/1.1.1d-fips zlib/1.2.11 libidn2/2.3.0 libpsl/0.21.0 (+libidn2/2.3.0) libssh/0.9.3/openssl/zlib nghttp2/1.40.0
 Data directory: /home/erbenton/boinc
 CUDA: NVIDIA GPU 0: GeForce RTX 2060 (driver version 440.59, CUDA version 10.2, compute capability 7.5, 4096MB, 3970MB available, 6739 GFLOPS peak)
 CUDA: NVIDIA GPU 1: GeForce GTX 1660 Ti (driver version 440.59, CUDA version 10.2, compute capability 7.5, 4096MB, 3972MB available, 5668 GFLOPS peak)
 OpenCL: NVIDIA GPU 0: GeForce RTX 2060 (driver version 440.59, device version OpenCL 1.2 CUDA, 5932MB, 3970MB available, 6739 GFLOPS peak)
 OpenCL: NVIDIA GPU 1: GeForce GTX 1660 Ti (driver version 440.59, device version OpenCL 1.2 CUDA, 5945MB, 3972MB available, 5668 GFLOPS peak)
 OpenCL CPU: pthread-Intel(R) Core(TM) i7-3960X CPU @ 3.30GHz (OpenCL driver vendor: The pocl project, driver version 1.4, device version OpenCL 1.2 pocl HSTR: pthread-x86_64-unknown-linux-gnu-sandybridge)
 Found app_info.xml; using anonymous platform
 [libc detection] gathered: 2.30, GNU libc
 Host name: erb1
 Processor: 12 GenuineIntel Intel(R) Core(TM) i7-3960X CPU @ 3.30GHz [Family 6 Model 45 Stepping 7]
 Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic popcnt tsc_deadline_timer aes xsave avx lahf_lm epb ssbd ibrs ibpb stibp tpr_shadow vnmi flexpriority ept vpid xsaveopt dtherm ida arat pln pts md_clear flush_l1d
 OS: Linux openSUSE: openSUSE Tumbleweed [5.4.12-1|libc 2.30 (GNU libc)]
 Memory: 31.29 GB physical, 2.00 GB virtual
6) Message boards : Number crunching : Need Help: trying to get 2 nvida cards runnng under linux (Message 2032169)
Posted 13 Feb 2020 by Profile Eric B
Post:
I'm running openSuse Tumbleweed. I had a nvidia-1660-ti as my primary gpu and it worked great. Now i have added a second nvidia card, an RTX 2060 and I swapped my video to this card. When i run boinc it sees and uses the RTX2060 but doesnt see the 1660ti. I am using the nvidia driver 440.31 and i tried re-installing that driver but it still doesn't work. Also nvidia-smi only sees one card (the 2060) but nvidia-settings sees both. Finally: CUDA Version is 10.2 according to nvidia-smi and my app_info.xml specifies the app setiathome_x41p_V0.98b1_x86_64-pc-linux-gnu_cuda101
lspci shows:
# lspci|grep VGA
         03:00.0 VGA compatible controller: NVIDIA Corporation TU106 [GeForce RTX 2060 Rev. A] (rev a1)
         05:00.0 VGA compatible controller: NVIDIA Corporation TU116 [GeForce GTX 1660 Ti] (rev a1)

I have use_all_gpu's set to 1 in cc_config
<cc_config>
 <options>
   <use_all_gpus>1</use_all_gpus>
 </options>
</cc_config>

boinc says:
12-Feb-2020 19:12:38 [---] CUDA: NVIDIA GPU 0: GeForce RTX 2060 (driver version 440.31, CUDA version 10.2, compute capability 7.5, 4096MB, 3970MB available, 6739 GFLOPS peak)
12-Feb-2020 19:12:38 [---] OpenCL: NVIDIA GPU 0: GeForce RTX 2060 (driver version 440.31, device version OpenCL 1.2 CUDA, 5932MB, 3970MB available, 6739 GFLOPS peak)
12-Feb-2020 19:12:38 [---] OpenCL CPU: pthread-Intel(R) Core(TM) i7-3960X CPU @ 3.30GHz (OpenCL driver vendor: The pocl project, driver version 1.4, device version OpenCL 1.2 pocl HSTR: pthread-x86_64-unknown-linux-gnu-sandybridge)
12-Feb-2020 19:12:38 [SETI@home] Found app_info.xml; using anonymous platform
12-Feb-2020 19:12:38 [---] [libc detection] gathered: 2.30, GNU libc


Any ideas?
TIA
7) Message boards : News : Low available work. (Message 2028684)
Posted 20 Jan 2020 by Profile Eric B
Post:
I'm curious to know what database program you use and what kind of database analysis tools you are using?
8) Questions and Answers : Web site : Why is the website sort feature so broken? (Message 2028503)
Posted 19 Jan 2020 by Profile Eric B
Post:
Its next to impossible to find the most recent threads on a given sub forum because the sort just doesn't work correctly.
Is it the website or is it my browser that's messed up? I've noticed this changes ordering at random times
For example, here is the first page of data taken from the "Number Crunching" sub forum sorted by newest first:
1 hour ago
3 hours ago
2 days ago
15 days ago
20 days ago
27 days ago
16 Dec 2019
9 Dec 2019
22 Mar 2019
53 minutes ago
3 hours ago
4 hours ago
5 hours ago
10 hours ago
11 hours ago
11 hours ago
17 hours ago
18 hours ago
1 day ago
1 day ago
1 day ago
9) Message boards : Number crunching : Is it possible to transfer tasks to another computer on my local network? (Message 2028332)
Posted 18 Jan 2020 by Profile Eric B
Post:
Hi, Thanks for responding !

Well, my ERB2 (I9-9900k) has 2 1660-ti's and it can do 2200 gpu tasks per day and around 1000 or so cpu tasks per day and it usually maxes out at about 6-800 tasks allotted per day so during an outage i very quickly run out of work on that guy
ERB1 (i7-3960x) with 1 1660-ti, is a bit slower and gets fewer tasks total, but after awhile (less than 24 hours usually) it too runs out
ERB3 is a laptop thats just dog slow, but oh well
ERB4 is an i7-2600k with a GTX 460 and its not real fast but for some reason it gets a ton of work to do
So i was looking to equalize things.
How does this work? How does the server decide how much work each PC can have?
Why does it seem to be so unbalanced in my case?
The Days Work is set the same for each PC "10 days plus 2 more"
I'll try what you said and change the numbers - see if it helps as of now it seems the slower the faster it gets work
Take this info for example :
Today (1/18/2020):                            
            Total   GPU  CPU 
  System     WUs    WUs 
----------------------------
   erb1      129     68   61
   erb2      115      0  115
   erb3      148      0  148
   erb4      297    150  147
----------------------------

1/15/2020:
            Total   GPU  CPU 
  System     WUs    WUs 
----------------------------
   erb1      389    218  171
   erb2      347    232  115
   erb3      198      0  198
   erb4      482    293  189
----------------------------

                              Average Credit per Host (as of today)
-----------------------------------------------------------------------------------
        erb1:              67,097  i7-3960X CPU @ 3.6GHz  w/ 1x GeForce GTX 1660 Ti
        erb2:             124,531  i9-9900K CPU @ 4.2GHz  w/ 2x GeForce GTX 1660 Ti
        erb3:               3,250  i7-4510U CPU @ 2.3GHz  w/ (No Usable GPU)
        erb4:              14,090  i7-2600K CPU @ 4.1GHz  w/ 1x GeForce GTX 460 v2
        User Avg Credit:  209,121
-----------------------------------------------------------------------------------


ps: I use OpenSuse Tumbleweed (the laptop(ERB3) is running Leap 15.1 i think) and have been using OpenSuse since before Leap came out, actually for about 10 years or so. I like it ok, I cant find a distro i like better tho.
10) Message boards : Number crunching : Is it possible to transfer tasks to another computer on my local network? (Message 2028327)
Posted 18 Jan 2020 by Profile Eric B
Post:
I have one computer that has lots and lots of cpu and gpu tasks but its a slow pc. I would like to take some of those tasks and move them over to my other pc which is (at least) an order of magnitude faster and at present is starving for work.
ps: I am on Linux exclusively, not Windows.
TIA
11) Message boards : Number crunching : The Server Issues / Outages Thread - Panic Mode On! (118) (Message 2024430)
Posted 22 Dec 2019 by Profile Eric B
Post:
Debugging the server is virtually impossible. If anyone wants to help.... The setiathome_server branch is at

https://github.com/BOINC/boinc/tree/setiathome_server/sched

Something goes wrong in the function SCHED_SHMEM::no_work.

bool SCHED_SHMEM::no_work(int pid) {
        if (!ready) return true;
        for (int i=0; i<max_wu_results; i++) {
            if (wu_results[i].state == WR_STATE_PRESENT) {
                wu_results[i].state = pid;
                return false;
            }
        }
        return true;
    }


This function works properly unless the requesting computer has anonymous platform apps, for which it always returns true. How could that be? I don't know despite additional 500 lines of debugging code. It's almost as if something else is pausing anonymous platform requests until the queue is empty. Well it's bed time now. :(


I guess my first question would be:
Is it returning true because of "!ready" ? or is it falling through and returning the bottom true.
If its falling through then either max_wu_results is less than zero or
wu_results[i].state is never equal to WR_STATE_PRESENT

Based on that analysis one can then decide what to look at next.
12) Message boards : Number crunching : Need help getting Inel GPU crunching (Message 2019494)
Posted 18 Nov 2019 by Profile Eric B
Post:
Yeh, I get your point about cooling. I also have a desktop i9-9900k with Intel gpu.
I installed a NVidia 1660 ti card but it would be useful to know how to setup that
intel gpu - I may never use the intel gpu but I'd still like to be able to
13) Message boards : Number crunching : Need help getting Inel GPU crunching (Message 2019387)
Posted 17 Nov 2019 by Profile Eric B
Post:
What about for newer Intel GPU's?
How does one setup for using those - for example the Intel UHD 630
14) Message boards : Number crunching : Need help getting Inel GPU crunching (Message 2019098)
Posted 15 Nov 2019 by Profile Eric B
Post:
.
Can anyone tell me how to get my GPU crunching Seti WU's?
openSUSE Leap 15.0 (x86_64) on kernel 4.12.14-lp150.12.79-default
CPU: Intel(R) Core(TM) i7-4510U CPU @ 2.00GHz
Installed Memory 8G
GPU: Intel Corporation Haswell-ULT Integrated Graphics Controller (rev 0b)
Its a laptop and even tho its performance is rather pitiful, I would still like to get the GPU up and running, MB's are already running on it without issue
15) Message boards : Number crunching : CL file missing (Message 2018683)
Posted 12 Nov 2019 by Profile Eric B
Post:
I'm on Linux Leap 15.1 and the nvidia driver was just updated.
Now the GPU tasks are postponed and then suspended due to a missing CL file.
What creates that file?
Its nvidia driver 390.132, the 390.129 driver was working great but this update to 132 broke things
16) Message boards : Number crunching : cuda 90 app not using as much GPU as it should? (Message 2018228)
Posted 9 Nov 2019 by Profile Eric B
Post:
Well, the same app AND the same WU - which brings me back to my original question -
is there some way to load the same WU on 2 different machines (multiple instances, 1 per boinc/seti cpu thread)
so I can properly compare things? I'd be using it as a benchmark WU. There needs to be safeguards so it is never
uploaded to a seti server
17) Message boards : Number crunching : cuda 90 app not using as much GPU as it should? (Message 2018166)
Posted 8 Nov 2019 by Profile Eric B
Post:
So after running awhile with the new settings I see the GPU's are working as expected but the CPU performance isn't. The i9 is lagging behind the i7 and the i9 has more threads. I gathered some data and it seems to confirm my visual observations. The average time to process MB WU's on the i9 is 4184 seconds each compared to the i7 at 3790. This isn't averaged over a lot of MB WU's but nevertheless it doesn't seem right to me, especially given that the i9 is running at 4.2Ghz vs the i7 at 3.7Ghz.

Average GPU time on ERB1: 100 GPU WUs - average of 73.3 seconds each
Average GPU time on ERB2: 100 GPU WUs - average of 61.4 seconds each

Average CPU time on ERB1: 19 CPU WUs - average of 3790.4 seconds each
Average CPU time on ERB2: 33 CPU WUs - average of 4184.6 seconds each

both PC's are using MBv8_8.05r3345_avx_linux64
I have two other MB apps I can try (if people think they would be better)
MBv8_8.22r3711_sse41_intel_x86_64-pc-linux-gnu
MBv8_8.22r4008_avx2_intel_x86_64-pc-linux-gnu

avx2 seems a step up from avx but probably it depends heavily on the app and if it can be taken advantage of or not.
18) Message boards : Number crunching : New PC not performing as well as expected (Message 2017903)
Posted 5 Nov 2019 by Profile Eric B
Post:
I only run run 1 gpu task at a time. I once 3experimentded with 2 tasks per gpu but it never panned out, performance just didnt add at all.
My 1660 ti's seems to be preforming quite well, it takes about 1:00 to 1:30 to do each task and actual RAC average is climbing dramatically since i made the changes above.
my GTX 1660 Ti's show (just checked):
i7-3960X
Average processing rate 2,303.31 GFLOPS
i9-9900k
Average processing rate 2,341.84 GFLOPS
19) Message boards : Number crunching : cuda 90 app not using as much GPU as it should? (Message 2017557)
Posted 2 Nov 2019 by Profile Eric B
Post:
I just switched one of my systems to cuda101 - if it pans out as a bit faster I'll switch the other also
I didn't know what to put for version_num so i just used 801 as before and it seems to work
Also I checked for trailing spaes
In vim use this
:highlight ExtraWhitespace ctermbg=red guibg=red
:match ExtraWhitespace /\s\+$/
and any trailing spaces show up in red
20) Message boards : Number crunching : cuda 90 app not using as much GPU as it should? (Message 2017333)
Posted 31 Oct 2019 by Profile Eric B
Post:
Whats the difference between
setiathome_x41p_V0.98b1_x86_64-pc-linux-gnu_cuda90
and
setiathome_x41p_V0.98b1_x86_64-pc-linux-gnu_cuda101

is one faster than the other?


Next 20


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.